A man named Victor cleaned my teeth. He was overly-detailed but kindly. I got the sense that he had been a dentist back in his home country and, like so many foreign physicians, had been forced to downgrade himself to a hygienist upon arriving in the States. When he took my medical history, he quickly and professionally informed me of the perils of dry mouth. (I take a lot of medications).

After roughly forty minutes, a woman named Samantha came in without announcing herself and looked at my teeth. She seemed too young to be a doctor, to be anything really. She was charismatic but didn’t ask me if I had any questions. I think I actually annoyed her when I didn’t let her take my sublingual bar out right then and there. “I have to get back to work,” I told her. When it was over, she left in a hurry and told me “Have fun at work.” (We both know she didn’t mean it).

Maybe when the maintenance guy never comes and the place blows sky high, Samantha and I can take the same canoe to heaven. Then she’ll have to forgive me. (Because I’ll do most of the paddling).

I recently watched Brandon Sanderson, one of the world’s most prolific living fiction authors, give a keynote on the use of AI in art. His discussion of the controversies surrounding AI in creative writing reminded me of NaNoWriMo’s deeply inflammatory statement on AI use in September 2024, which is often cited (albeit incorrectly) as one of the reasons for the organization’s shuttering announced in March 2025. The statement was as (roughly) as follows:

“[NaNoWriMo believes that] categorical condemnation of artificial intelligence has classist and ableist undertones, and that questions around the use of AI tie in to questions about privilege.” (Wired).

As someone who has ADHD and has always struggled with writing anything in excess of 10,000 words, I’ve played around with the idea of using AI to execute creative writing projects, though in a very specific, narrow sense. (Spoiler: it’s bad).

Still, I haven’t been able to get NaNoWriMo’s statement or the backlash towards it out of my head. This is mostly because I don’t understand what, if anything, is being said in this all-too-brief sentence. On the surface, NaNoWriMo’s statement on AI has the appearance of being a vapid, virtue-signaling political jab. Statements that take the form “X is Y, Y is bad, therefore X is bad” are not meant to be complex arguments or meaningful statements on moral values; they’re a shorthand application of guilt by association, which at best stands in for a longer chain of reasoning, and at worst obscures the fact that no such reasoning exists.

I want to try to understand this statement in good faith, meaning I’m making the assumption that there is at least some thought behind this statement, that the chain of reasoning referenced by the author exists somewhere in the ether, to say nothing of its quality or completeness. My goal here is to imagine what that chain of reasoning might have been, if and where it went wrong, and why that matters for the future of writing.

The way I see it, this statement makes two claims. The first is that a categorical condemnation of artificial intelligence (in creative writing) is classist. The second is that a categorical condemnation of artificial intelligence (in creative writing) is ableist. Let’s discuss each of these claims in turn.

Classism has become a catch-all term used to describe everything from corporate dress codes to the way bus routes are mapped through cities. In its simplest form, classism is enacted stigma against members of a particular socioeconomic group (“class”). This is almost exclusively done by higher classes to lower classes, however the inverse does, in theory, exist. (Although the notion of anti-rich bias is laughable in an economic system where money is exchangeable for pretty much any form of power or resource).

Taking this definition, it would seem the author of the Statement is implying that a condemnation of AI favors the higher classes at the expense of lower classes. One might think it would be the opposite — after all, good AI services are paywalled, and it would seem that those who can afford to integrate ChatGPT into their writing practices have a leg up on those without machine assistance. On the surface, it would seem that outlawing the use of AI in a writing space might actually level the playing field between classes by ensuring that everyone has access to the same set of tools.

The argument that a no-AI policy actually privileges the upper classes is somewhat more circuitous. It can be reasoned that AI use by lower-class people enables them to produce writing which is either higher quality or more prolific than would otherwise be possible due to socioeconomic constraints. Wealthy writers often have expensive, high-quality educations and can afford to spend more time writing than people who might need to work multiple jobs to make ends meet. This is to say nothing of things like writer’s retreats, paid editing services, access to professional networks, etc.

In this resource context, it may be possible to argue that AI (which is typically free and easy to access, though one does start to run into issues when generating output at large scales) may help writers with access to fewer resources produce writing which can compete with their upper-class peers.

There’s a few problems with this approach. First, it implicitly assumes the goal of writing is the final (commercial) product, rather than the process itself. As Sanderson argues, the purpose of writing is not necessarily what’s at the end, but the process of getting there. It is through trial and error, pouring over sentences again and again, and failure that one learns how to become a writer, and this becoming is the “good thing” of writing, that which makes the act of writing beneficial to the human who does it. AI may be a shortcut to the product, but in a paradigm where the act of writing itself is both what’s worthwhile and a luxury of time and resources, lower-class people still get the short end of the stick. Second, this reasoning makes less-than-favorable assumptions about lower-class peoples’ ability to write or produce high quality writing in spite of their lower-class status. Good writing is not merely the product of time and education; a person must first and foremost have something to say, a story to tell, before they can write it. (We likely all know someone in our lives who, despite having access to seemingly every resource under the sun, still bumbles through most of the time and never manages to produce anything interesting, literary or otherwise).

When it comes to the act of producing (again, creative) writing, it would seem that allowing AI may, under a very particular set of assumptions, enable someone from a lower class to compete (commercially) with an upper-class person who has access to a superior set of resources if said upper-class person did not also have access to AI. And herein lies the problem. As with most tools, there is a learning curve and/or use cost, and a person with access to wealth has more time, resources, and computational hardware to make better use of an otherwise publicly-available AI tool (again, assuming commercial output is the goal).

More generally, any argument which assumes access to some resource will level the playing field between groups must assume that the underprivileged group has differential access to and/or use of the resource compared to the privileged group. This is the classic issue of equality vs. equity — the person with the most resources will always win the race unless you actively move the underprivileged person ahead. Given this, the claim that categorically banning AI use is somehow classist is deeply flawed, since allowing AI at best does nothing to change the relative position of the two classes, and at worst widens the gap between them.

We can thus reject the first claim of the Statement, and move on to the second.

The notion that banning AI is ableist has more heft to it immediately. If we define ableism as enacted stigma against people with disabilities (including accommodations for those disabilities), then forbidding AI use in creative writing seemingly forbids a certain kind of person from participating who otherwise could not participate but for the use of AI. Essentially, if AI is treated as an accommodation, then disallowing it is ableist.

Now the question remains, who needs AI to write? The answer, simply, is someone. It might be a person with ADHD who struggles to get started or finish their thoughts. It might be a person with limited cognitive abilities who cannot focus on a task for long periods of time. It might not be everyone, and it might not even be common, but asking who needs AI to write is in some ways akin to asking who needs crutches to walk. We can ask questions about the ethics of AI as an accommodation, or whether the use of AI robs certain disabled people from the experience of writing as described by Sanderson, but we cannot categorically deny it when it is employed as an accommodation, at least not without offering an alternative. (For instance, it may not be possible to offer someone a particular kind of mobility assistance, but some mobility assistance must be offered if lacking access to such assistance forbids the person from participating in an “accessible” group endeavor).

As with class, this line of reasoning makes less-than-favorable assumptions about disabled peoples’ ability to write without AI, but this opens up a disability politics question more than anything else. The disabled writers who say “I can write without AI, how dare you say I can’t!” are fundamentally disabled in a different way than those disabled writers who rely on AI as part of their practices. Then the question becomes who gets to use the term “disabled,” and the ingroup fighting devolves from there. (Note too that, in the social model of disability, we could reframe being lower-class as a disability, and hence arrive at the same conclusion with regards to class if the issue at hand is access rather than overall outcomes).

We have thus broken up the kinds of discrimination alleged by the Statement and gotten two very different results. The claim that banning AI is classist is nonsensical, but the claim that banning AI is ableist is a fairly straightforward disability justice take (once one looks past its immediate political implications for high-functioning disabled people who can write without AI assistance). I suspect that if ableism had preceded classism in the Statement, it may have been better (but still not well-) received. Nevertheless, it doesn’t seem nearly controversial enough to cause the fuss it did, and certainly wasn’t controversial enough to singlehandedly bring down NaNoWriMo as an enterprise. The culprit there, as in so many organizational failings, was the inability to orchestrate resources across complex systems. Unfortunately for NaNoWriMo, “longstanding cross-systems liquidity issues” is decidedly less punchy than “go woke go broke.”

In high school, I had the choice of taking ancient Greek or an additional science. My mother’s friend whose wife went to Yale told her that Greek was a waste of time, which naturally meant that this was and had always been an immutable truth of the universe.

Lo and behold, I did not take Greek.

But I remember wanting to take it.

I think what I loved most about it at the time was that even then I knew ancient Greek would be completely and utterly useless. All of my friends who were “bad influences” were taking it. It was flashy but not substantive, a frivolous choice made by unscrupulous young men who did not have their futures in mind. I just thought it was interesting.

I liked the funny letters.

As someone who’s spent significant time working with older adults, there’s an ethical dilemma I encounter frequently that goes something like this: “my loved one has dementia, and what they want now is different from what they wanted before they had dementia. Which set of wishes do I honor?”

There’s lots of good thought on this subject. My own contribution is an attempt to offer a concise set of “best practices” about how we might go about determining whether we honor past or present wishes when an individual’s decision-making capacity may be compromised due to dementia symptoms.

Many people don’t understand that individuals living with dementia can still have values, act on those values, and change their values, even in the late stages of cognitive decline. It remains important to make a careful study of an individual’s values trajectory over time to create the best possible chance of honoring what their “authentic” values are in a given moment.

Respecting the wishes of a person living with dementia compels us to ask essential questions about the meaning and continuity of personal values. Presumably, our goal in these situations is to honor the authentic wishes of the person living with dementia. We might think of authentic wishes as those decisions which arise from reflection on one’s values (Menzel & Steinbock, 2013). Reflection on one’s values means that a person makes their decision with full capacity to appreciate:

(1) their values,

(2) how their decisions relate to those values, and

(3) the consequences of those decisions.

The crucial issue is that dementia can lead to false beliefs and/or cognitive deficits, which may directly interfere with (1), (2), or (3).

First and foremost, a person can forget their values or develop false beliefs about what their values actually are. Imagine a person who, prior to any indicators of cognitive decline, expresses their desire to die once certain they lose certain capacities (such as no longer recognizing their family) in accordance with their values. This person may then go on to develop the false belief in dementia that they value being alive at any cost. When they express this value, it can be unclear to caregivers and other stakeholders whether this is a legitimate expression of a new value or the illegitimate expression of values based in false beliefs.

Values change naturally as we experience new things and reflect on our lives in new ways. Individuals with dementia, even advanced dementia, can retain the capacity to reflect meaningfully on their experiences, thereby producing (and expressing) authentic values which differ from previous ones (Wilkins, 2017). Were dementia or another change in capacity not present, we would unquestioningly honor the person’s most recent set of values in decision-making. However, this is precisely the issue dementia presents: we do not know for what periods of time (if ever) an individual loses their ability to develop and/or express their authentic values. If we are uncertain what a person’s most recent authentic values are, we cannot honor them.

In a second scenario, individuals may be aware of their authentic values but lose the capacity to relate their decisions to those values. Imagine a person who authentically values being kind to others but cannot comprehend how continuing to live with dementia-related aggression could result in unkindness.

Additionally, we might consider individuals who are aware of their authentic values and can relate their decisions to them, but cannot appreciate the consequences of those decisions. Think of someone who adheres to a particular faith tradition and whose authentic values are meaningfully related to the decision to die a certain kind of death, but cannot appreciate the consequences of that death in the context of their faith tradition due to cognitive decline.

While we can never be fully certain if new statements about values from persons experiencing dementia are fully legitimate, there exist numerous clinical tools which can serve as heuristics in helping us with our decision-making. For example, the MacArthur Competence Assessment Tool (MacCAT) includes questions which can assist providers in determining to what extent a person has capacity to make legitimate statements about their wishes. Questions probe patients on their reasoning to test their ability to connect values to decisions. Other questions evaluate patients’ comprehension of consequences by asking them to describe the outcomes or risks associated with a particular decision.

Despite the utility of the MacCAT and similar tools, it is well-documented that individuals living with Alzheimer’s and other types of dementia can have periods of “lucidity” in which they seemingly return to their old selves, interspersed by equally potent periods of the cognitive symptoms described above (Jacobbi, 2024). Capacity assessments are essential, but what are we to do if the results of these instruments contradict one another, are inconclusive, or fluctuate so rapidly as to be effectively meaningless, rendering us unable to assess someone’s cognitive ability from moment to moment, much less “on average”?

If the individual in question did not have cognitive challenges, we would default to their most recent set of expressed values. Therefore, it is important for caretakers and other stakeholders to consistently reassess a patient’s capacity and the value statements they make during that state. Legitimate changes in authentic values can occur late in life and in individuals with dementia; producing a strong record of values statements made during high-capacity moments is the best way to furnish evidence of an individual’s true wishes. This means asking questions about values and assessing capacity early and often.

References

Alzheimer’s Stages — Early, Middle, Late Dementia Symptoms | alz.org. (n.d.). Alzheimer’s Association. Retrieved July 6, 2025, from https://www.alz.org/alzheimers-dementia/stages

Appelbaum, P. S. (2007). Assessment of Patients’ Competence to Consent to Treatment. N Engl j Med.

Appelbaum, P. S., & Roth, L. H. (1983). Patients who refuse treatment in medical hospitals. JAMA, 250(10), 1296–1301.

Dementia UK. (2023). False beliefs and delusions in dementia. Dementia UK. https://www.dementiauk.org/information-and-support/health-advice/false-beliefs-and-delusions-in-dementia/

Grisso, T., Appelbaum, P. S., & Hill-Fotouhi, C. (1997). The MacCAT-T: A clinical tool to assess patients’ capacities to make treatment decisions. Psychiatric Services (Washington, D.C.), 48(11), 1415–1419. https://doi.org/10.1176/ps.48.11.1415

Jacobbi, V. (2024, March 4). Moments of clarity in the fog of dementia. Mayo Clinic News Network. https://newsnetwork.mayoclinic.org/discussion/understanding-lucid-episodes-in-dementia/

Menzel, P. T., & Steinbock, B. (2013). Advance directives, dementia, and physician-assisted death. The Journal of Law, Medicine & Ethics: A Journal of the American Society of Law, Medicine & Ethics, 41(2), 484–500. https://doi.org/10.1111/jlme.12057

Scott, D. (2008). Toolkit for Primary Care: Capacity Assessment. Primary Care.

Treem, J. (2023). Medical Aid in Dying: Ethical and Practical Issues. Journal of the Advanced Practitioner in Oncology, 14(3), 207–211. https://doi.org/10.6004/jadpro.2023.14.3.5

Wendler, D., & Rid, A. (2011). Systematic review: The effect on surrogates of making treatment decisions for others. Annals of Internal Medicine, 154(5), 336–346. https://doi.org/10.7326/0003-4819-154-5-201103010-00008

Wilkins, J. M. (2017). Dementia, Decision Making, and Quality of Life. AMA Journal of Ethics, 19(7), 637–639. https://doi.org/10.1001/journalofethics.2017.19.7.fred1-1707

(Throughout this article, I use the term “AI” in its colloquial, general sense to refer to large-language models oriented towards producing written responses, of which ChatGPT is perhaps the most notable example. See here for a more formal definition).

Introduction

First, some preliminaries. I wanted to write a novel, which, in this project, I defined as a continuous work of fiction at least 50,000 words in length (in line with the NaNoWriMo definition). From the beginning, I wanted the project to be AI-collaborative, not AI-driven. To this end, I did not use AI for any of the big creative decisions. I created the characters, the plot beats, themes, tensions, etc. My idea was to have the AI execute this vision, to do that most laborious task of translating a rough schema with good ideas into a full-length, polished text. In essence, I was hoping AI would do the “gruntwork” of writing, leaving the creative part (the fun stuff) to me.

There are many ways one can go about using AI to write a novel. I used a software called Novelcrafter (not sponsored), which has some basic quality-of-life improvements for novelists over conventional word-processing software. This includes the ability to efficiently organize things into scenes, chapters, and acts, as well as creating a codex of key characters/objects. For example, while I was writing, I could click on a character’s name, and see a description of them, their relationship to other characters, etc. These features alone were useful to me as a writer, before any AI.

Novelcrafter’s specialty, however, is AI-integration. Unlike other services, Novelcrafter does not have its own “in house” AI. Instead, it connects you to a website called OpenRouter, which is effectively an AI switchboard. You can choose from a variety of different models (both free and paid) to accomplish writing tasks. In this project, I mainly utilized GPT-4o (as it stood in late May/early June of 2025). I did have to pay for it, but $5 worth of “credits” on OpenRouter got me way more than what I needed (after this entire endeavor, I had only used up about 50 cents’ worth, though I imagine someone using AI more intensively could go through their credits faster). I tried experimenting with some other models besides GPT-4o, but none performed anywhere near as well as what I needed (which isn’t saying much — again, nothing sponsored/endorsed here).

Getting Into It

My collaboration with AI was generally as follows: I would give the AI a scene outline and ask it to write it out — this means that the AI would generate the entire scene based on my notes, including references to characters in the novel’s codex, previously written sections, etc. (one of the advantages of Novelcrafter is that you can efficiently prompt the AI with codex entries rather than writing out specific references each time). This process was always iterative. I would have the AI produce and/or rewrite the same text multiple times, often making my own edits/tweaks to the base text between prompts or fully scrapping and rewriting some parts before giving it back to the AI again, along with at least two thorough passes of my own editing once I had gotten what I wanted from the model.

I would attempt to prompt the AI into generating the same text in a different way by giving instructions such as “be less dramatic here,” “shorten this,” or “don’t mention X.” My most effective and favored revision prompt was “rewrite this like Hemingway would have.” I have no particular affiliation for Hemingway, but I found that prompting it this way gave the most usable text as output (superior even to similar instructions like “be short, be concise, eliminate unnecessary details”). I think one of the reasons this worked so well is that large language models are, at their most fundamental, probability-driven imitators. I can say “be short” but saying “be short like Hemingway” immediately cues the algorithm to use Hemingway’s corpus as a data source (which is full of not only short sentences, but good short sentences). I found myself using this prompt frequently given that the base voice of ChatGPT at the time I was writing it tended to be overly flowery and descriptive in ways that didn’t make sense. Asking the model to write with only the bare bones of a scene (like Hemingway) produced the most useful thing to work with, if not the finished product. This brings me to my next point, which is:

AI is a tool, not a coauthor

I began with the lofty idea that an AI would faithfully execute my artistic vision, filling in all the gaps that my ADHD brain just could not be bothered with. To this day, the main reason I have not produced a full-length novel despite having drafted many detailed, novel-length outlines is mainly my brain’s absolute refusal to work with the same topic for more than roughly 2,000 words. I thought AI would close that gap, doing the dry, mechanical work of composition that was necessary to move the plot along, adding on a flourish or two, and leaving the rest, the literary, to me. This did not happen.

The AI model I used was incredibly repetitive, even when given different prompts. I would ask it to describe something — say, an idyllic Italian town — in different ways, and it would change a few words out each time, but the underlying image was functionally the same. I spent the better part of a half-hour trying to get it to not describe a character as being “enveloped” by darkness. With every iteration, it found a different way of saying the same thing (“enshrouded,” “suffocated,” “pressed by”). The model could not give up the base idea it had about a scene, even when explicitly asked to do so (“do not describe character X using any concept of darkness, being covered by or enveloped in darkness, and related ideas”).

It’s possible that I could have worked around this; many AIs have parameters which can be altered by a user to change the predictability of the output. In technical terms, this is called adjusting the temperature of the model. Higher temperatures mean more randomness in the output, lower temperatures mean more consistency. A model temperature of absolute 0 (in whatever scale) should theoretically produce the same output in response to the same input every time. I did not adjust the temperature on the model itself during this project, but I find it noteworthy that instructions to describe something explicitly without mention of certain words or phrases often failed spectacularly. The model was too rigidly tied to its own internal image of the “idyllic Italian town” or similar to meaningfully deviate from it, even when prompted to do so. (I’m not a technical person by any means, but my guess is that in the probability space of the model, the words and phrases related to the ideas that I was working with were clustered too close together — it couldn’t mention A and not mention B, much less reach out for C, even when heavily weighting for C in my prompting or explicitly forbidding it from using B).

This is also to say nothing of the AI’s repeated use of the same descriptors across instances, even when explicitly prompted to not repeat itself with the context of prior text. As you might imagine, this only got more difficult to deal with as I went further along in the drafting process. It got to the point where the AI would use the same descriptors for scenes or characters that really had no similarity to one another. When given complex or abstract things to describe, its output only became worse in both quality and repetitiveness.

Writing is Rewriting, for Man and Machine

Any productivity gains were often outweighed by the significant amount of rewriting that I had to do on the AI-generated text, both with and without the AI’s help. For example, I would often give slight variations of the same prompt two or three times before getting text which I considered passable (which I would then manually edit and touch up at least two more times). Or I would ask it to produce five sentences, only two of which ended up usable, meaning either I had to write the other three myself or generate a new prompt to produce them, etc. With each prompting, it took me so long to corral the model into producing something which, at the end of the day, was nothing a well-read high-school senior couldn’t write in 10 minutes.

I will say this — the AI was extremely helpful at getting over the hump of starting or continuing writing. Instead of staring at a blank page, I could have the AI produce a page of writing — it wasn’t good writing, by any means, but at least then I’d have something to work with. Editing writing, even significantly, is almost always easier than producing it all from scratch. There were also moments when I wasn’t sure how to navigate a transition scene (say, a character leaving a cafe and walking out into the street). Asking the AI to write these sections produced the skeleton, the base mechanics of what I wanted, from which I could modify as needed to suit the particulars of the setting, the characters, etc. As someone who has trouble with task initiation, this led me to write both more and more often than I otherwise would have, but it wasn’t magic. If I really didn’t want to write on a particular day, having an AI paragraph or page to get me started wouldn’t make a difference.

Conclusion

In the end, despite having a full outline, I only got through about 10,000 words before abandoning the project. I had largely lost interest in the subject matter, the AI wasn’t giving me the productivity I had hoped for, and the novelty of an AI-collaborative novel had thoroughly worn off, replaced instead by the drudgery of prompt engineering to mediocre results. I’m glad I at least attempted this work — it taught me a lot about myself as a writer, along with the limitations of AI in creative writing. I now know how to use AI to augment, rather than replace, my writing process: the technology really shines when it comes to breaking through writer’s block and changing the style (but not substance) of texts. It cannot replace human ideas by any means, and when it is asked to write, produces what feels like (and largely is) the statistical average of its corpus, plus or minus a few terms. I had hoped that AI would close that dreaded gap between outline and manuscript where so many writing projects fail, but as things stand now, that’s not the case. It’s more akin to a random number generator, a fancy spellcheck, or an enhanced form of Googling than it is an actual (co)author, despite what its creators may claim.

As someone who works a lot with older adults, one of the things I’m asked about most often is Alzheimer’s disease (AD): who gets it, when do they get it, why do they get it, how can it be identified, prevented, reversed, etc. There are good answers to these questions. But despite working with AD patients on a daily basis and knowing a lot about how to support people with AD and their caregivers, I was plagued by a sense that I didn’t have an understanding of what AD really was. Like, when we talk about Alzheimer’s, what is it that we’re actually referring to? This led me down a rabbit hole of trying to understand how AD is defined and diagnosed in the strictest, most concrete terms. After substantial research, I finally began to wrap my head around what AD is in a biological sense and in a behavioral sense, and why diagnosing it (or anything that arises from the interplay of biology and behavior) remains challenging. This article is the result of that research, an attempt at an overview of why diagnosing Alzheimer’s disease is just so damn complicated.

Alzheimer’s disease (AD) is defined primarily by two physiological changes which occur in the brain:

1. Amyloid-beta plaque buildup (APB)
2. Tau protein tangles (TPT)

AD can only be definitely diagnosed after death via autopsy, when a physician is able to confirm the presence of both APB and TPT. However, doctors have tests which can reasonably (but not infallibly) predict the presence of APB or TPT in someone’s brain while they’re alive. While these tests alone are not sufficient for a definitive Alzheimer’s diagnosis, they can be used in conjunction with other biological or behavioral criteria to inform a probable diagnosis of AD in a living patient.

Here it is important to note that while APB and TPT define AD, they do not necessarily cause AD. For example, some people who develop amyloid-beta plaque buildup do not develop symptoms of AD. (Moreover, if APB alone was sufficient to cause AD, then removing amyloid-beta plaque and/or preventing its accumulation should prevent AD. However, drugs which effectively remove or prevent amyloid-beta plaque buildup in the brain do not improve the symptoms of AD in patients already exhibiting cognitive decline (Digma et al., 2024)).

Broadly speaking, there are two kinds of physiologic tests used to evaluate the presence and concentration APB and TPT in someone’s brain and thus provide a probable diagnosis of AD in living patients.

The first test is a positron emission tomography (PET) scan. It works by a physician injecting a radioactive drug into the body, which is designed to bind with either APB or TPT. Once the drug binds it stays with the APB or TPT and begins to decay, emitting positrons. Positrons are very small, positively-charged subatomic particles. These particles can be picked up by a scanner, and used to create an image which shows the relative location and concentration of APB or TPT in someone’s body. By comparing the image from the PET scan of a potential AD patient to an image of a physiologically-healthy brain which has undergone the same procedure, a clinician can determine whether there is sufficient buildup of APB or TPT in the right locations to give a probable diagnosis of AD. PET scans which are specifically designed to target APB have a higher diagnostic accuracy than those targeting TPT. This is often attributed tothe fact that TPT buildup can also be associated with other diseases which create symptoms similar to AD (Gao et al., 2018).

The second kind of test is a blood test. A blood test looks for proteins associated with the buildup of APB or TPT in the brain. Notably, unlike PET scans, blood tests which evaluate only APB and blood tests which evaluate only TPT have similar accuracy in correctly diagnosing AD. Blood tests are now considered equivalent to PET scans in diagnosing AD, and are often preferred given their lower cost and greater accessibility in primary care settings (Reynolds, 2024).

In addition to biological indicators of AD, clinicians often use studies of cognitive impairment to make a probable diagnosis. These cognitive tests vary widely in what they test, their length, and how or when clinicians decide to administer them. The body of available tests and diagnostic algorithms in this regard is so great that it cannot be reviewed adequately here. Generally however, these tests involve the assessment of an individual’s short-term/working memory and/or language abilities (Alzheimer’s Disease Fact Sheet, 2023).

There is no agreed-upon algorithm or procedure to definitively diagnose AD in living patients. A joint commission of the National Institute on Aging and the Alzheimer’s Association considers the abnormal presence of certain biomarkers to be sufficient for a diagnosis in a living person who is exhibiting symptoms (Jack Jr. et al., 2024). However, in the absence of clear biomarkers, AD is diagnosed based on a mix of non-preferred biomarkers, cognitive tests, family medical histories, and clinician judgement (Jack Jr. et al., 2024).

Like many other neurodegenerative diseases, Alzheimer’s exists on a spectrum, and there is no direct or systematic correlation between a certain biomarker value and/or test score with a particular stage/severity of the disease. A person could have abnormal biomarkers and not exhibit any symptoms of AD or vice versa. Perhaps most famously in a study conducted on nuns, there were cases in which subjects exhibited all the symptoms of AD but minimal to none of the associated brain pathology or moderate to severe brain pathology with none of the symptoms (Iacono et al., 2009). This lack of alignment between physiology and symptomatology is one of the central reasons why diagnosing AD remains difficult in living patients.

Another reason diagnosing AD in living patients continues to be challenging is the potential for multiple or “mixed” dementias. These are cases in which multiple disease pathologies are responsible for a person’s cognitive decline. For example, vascular dementia is dementia caused by an interruption of blood flow to the brain and one of the most common causes of neurodegenerative disease outside of AD (Sanders et al., 2024). A patient may experience vascular dementia and AD simultaneously, but the vascular dementia symptoms may mask the onset of symptoms which are otherwise attributable to AD. In this case, it is likely that the patient’s AD would only be discovered upon autopsy, if they are autopsied at all. To distinguish between multiple dementia presentations, a clinician may attempt to treat all possible dementias simultaneously, or see which treatment approaches produce the best outcomes. For example, mixed dementia which is primarily vascular in character may be slowed (but not reversed) with lifestyle improvements, which is generally not the case with AD (Sanders et al., 2024).

Diagnostic accuracy for AD varies widely depending on the specialty of the physician doing the diagnosis and whether biomarkers, clinical presentations, or a combination of both are considered. Generally, biomarkers remain the largest determinant in a correct diagnosis, but must always be supplemented by clinician judgement. (Recall that the presence of biomarkers alone may not indicate AD in an asymptomatic patient) (Reynolds, 2024).

References

Alzheimer Society of Canada. (n.d.). How Alzheimer’s disease changes the brain. Alzheimer Society of Canada. Retrieved November 29, 2024, from https://alzheimer.ca/en/about-dementia/what-alzheimers-disease/how-alzheimers-disease-changes-brain

Alzheimer’s Association. (2024a). 10 Early Signs and Symptoms of Alzheimer’s and Dementia. Alzheimer’s Disease and Dementia. https://alz.org/alzheimers-dementia/10_signs

Alzheimer’s Association. (2024b). Mixed Dementia. Alzheimer’s Disease and Dementia. https://alz.org/alzheimers-dementia/what-is-dementia/types-of-dementia/mixed-dementia

Beach, T. G., Monsell, S. E., Phillips, L. E., & Kukull, W. (2012). Accuracy of the Clinical Diagnosis of Alzheimer Disease at National Institute on Aging Alzheimer’s Disease Centers, 2005–2010. Journal of Neuropathology and Experimental Neurology, 71(4), 266–273. https://doi.org/10.1097/NEN.0b013e31824b211b

Chávez-Fumagalli, M. A., Shrivastava, P., Aguilar-Pineda, J. A., Nieto-Montesinos, R., Del-Carpio, G. D., Peralta-Mestas, A., Caracela-Zeballos, C., Valdez-Lazo, G., Fernandez-Macedo, V., Pino-Figueroa, A., Vera-Lopez, K. J., & Lino Cardenas, C. L. (n.d.). Diagnosis of Alzheimer’s Disease in Developed and Developing Countries: Systematic Review and Meta-Analysis of Diagnostic Test Accuracy. Journal of Alzheimer’s Disease Reports, 5(1), 15–30. https://doi.org/10.3233/ADR-200263

Cummings, J. L., Isaacson, R. S., Schmitt, F. A., & Velting, D. M. (2015). A practical algorithm for managing Alzheimer’s disease: What, when, and why? Annals of Clinical and Translational Neurology, 2(3), 307–323. https://doi.org/10.1002/acn3.166

Custodio, N., Montesinos, R., Lira, D., Herrera-Pérez, E., Bardales, Y., & Valeriano-Lorenzo, L. (2017). Mixed dementia: A review of the evidence. Dementia & Neuropsychologia, 11(4), 364–370. https://doi.org/10.1590/1980-57642016dn11-040005

Digma, L. A., Winer, J. R., & Greicius, M. D. (2024). Substantial Doubt Remains about the Efficacy of Anti-Amyloid Antibodies. Journal of Alzheimer’s Disease, 97(2), 567–572. https://doi.org/10.3233/JAD-231198

Gao, Y.-L., Wang, N., Sun, F.-R., Cao, X.-P., Zhang, W., & Yu, J.-T. (2018). Tau in neurodegenerative disease. Annals of Translational Medicine, 6(10), 175. https://doi.org/10.21037/atm.2018.04.23

Hahr, J. Y. (2015). Physiology of the Alzheimer’s disease. Medical Hypotheses, 85(6), 944–946. https://doi.org/10.1016/j.mehy.2015.09.005

Iacono, D., Markesbery, W. R., Gross, M., Pletnikova, O., Rudow, G., Zandi, P., & Troncoso, J. C. (2009). The Nun Study. Neurology, 73(9), 665–673. https://doi.org/10.1212/WNL.0b013e3181b01077

Jack Jr., C. R., Andrews, J. S., Beach, T. G., Buracchio, T., Dunn, B., Graf, A., Hansson, O., Ho, C., Jagust, W., McDade, E., Molinuevo, J. L., Okonkwo, O. C., Pani, L., Rafii, M. S., Scheltens, P., Siemers, E., Snyder, H. M., Sperling, R., Teunissen, C. E., & Carrillo, M. C. (2024). Revised criteria for diagnosis and staging of Alzheimer’s disease: Alzheimer’s Association Workgroup. Alzheimer’s & Dementia, 20(8), 5143–5169. https://doi.org/10.1002/alz.13859

Lewczuk, P., Matzen, A., Blennow, K., Parnetti, L., Molinuevo, J. L., Eusebi, P., Kornhuber, J., Morris, J. C., & Fagan, A. M. (2016). Cerebrospinal Fluid Aβ42/40 Corresponds Better than Aβ42 to Amyloid PET in Alzheimer’s Disease. Journal of Alzheimer’s Disease, 55(2), 813–822. https://doi.org/10.3233/JAD-160722

Mayo Clinic Labs. (n.d.). Mayo Clinic Test Definition: C2AD2. Mayo Clinic Labs. Retrieved November 29, 2024, from https://www.mayocliniclabs.com/api/sitecore/TestCatalog/DownloadTestCatalog?testId=621652

National Health Service. (2017, October 23). Vascular dementia. National Health Service. https://www.nhs.uk/conditions/vascular-dementia/

National Institute on Aging. (2023, April 5). Alzheimer’s Disease Fact Sheet. National Institute on Aging. https://www.nia.nih.gov/health/alzheimers-and-dementia/alzheimers-disease-fact-sheet

National Institute on Aging. (2024, July 11). Alzheimer’s Caregiving: Managing Personality and Behavior Changes. National Institute on Aging. https://www.nia.nih.gov/health/alzheimers-changes-behavior-and-communication/alzheimers-caregiving-managing-personality-and

Palmqvist, S., Tideman, P., Mattsson-Carlgren, N., Schindler, S. E., Smith, R., Ossenkoppele, R., Calling, S., West, T., Monane, M., Verghese, P. B., Braunstein, J. B., Blennow, K., Janelidze, S., Stomrud, E., Salvadó, G., & Hansson, O. (2024). Blood Biomarkers to Detect Alzheimer Disease in Primary Care and Secondary Care. JAMA, 332(15), 1245–1257. https://doi.org/10.1001/jama.2024.13855

Reynolds, S. (2024, August 14). Accurate blood test for Alzheimer’s disease. National Institute on Aging. https://www.nia.nih.gov/news/accurate-blood-test-alzheimers-disease

Sabbagh, M. N., Lue, L.-F., Fayard, D., & Shi, J. (2017). Increasing Precision of Clinical Diagnosis of Alzheimer’s Disease Using a Combined Algorithm Incorporating Clinical and Novel Biomarker Data. Neurology and Therapy, 6(Suppl 1), 83–95. https://doi.org/10.1007/s40120-017-0069-5

Sanders, A. E., Schoo, C., & Kalish, V. B. (2024). Vascular Dementia. In StatPearls. StatPearls Publishing. http://www.ncbi.nlm.nih.gov/books/NBK430817/

Snowdon, D. A. & Nun Study. (2003). Healthy aging and dementia: Findings from the Nun Study. Annals of Internal Medicine, 139(5 Pt 2), 450–454. https://doi.org/10.7326/0003-4819-139-5_part_2-200309021-00014

Turk, K. W., Vives-Rodriguez, A., Schiloski, K. A., Marin, A., Wang, R., Singh, P., Hajos, G. P., Powsner, R., DeCaro, R., & Budson, A. E. (2022). Amyloid PET ordering practices in a memory disorders clinic. Alzheimer’s & Dementia: Translational Research & Clinical Interventions, 8(1), e12333. https://doi.org/10.1002/trc2.12333

Unterrainer, M., Eze, C., Ilhan, H., Marschner, S., Roengvoraphoj, O., Schmidt-Hegemann, N. S., Walter, F., Kunz, W. G., Rosenschöld, P. M. A., Jeraj, R., Albert, N. L., Grosu, A. L., Niyazi, M., Bartenstein, P., & Belka, C. (2020). Recent advances of PET imaging in clinical radiation oncology. Radiation Oncology, 15(1), 88. https://doi.org/10.1186/s13014-020-01519-1

Narcissism isn’t real. This is not to say there aren’t people who are profoundly self-centered, arrogant, or have a delusional sense of self importance. Rather, the category of the narcissist is an invention of modern psychiatry. But most mental health diagnoses are just lines in the sand drawn by psychiatrists, subjective categories that facilitate treatment planning. Why focus on narcissism?

I chose to write about narcissism because there seems to be immense popular interest in how to respond to so-called “narcissists” in everyday life. A cursory search returns articles like “Things Narcissists Say in An Argument and How to Respond,” “10 Phrases to Disarm a Narcissist,” or my personal favorite: “How to Destroy a Narcissist.” These articles take for granted the idea that there exists some real-world category, some actual phenomena, of a person that is definitely separable from the rest of humanity based on their behavior. And while I certainly think we can speak of a narcissistic person, I want to problematize the diagnostic category of “the narcissist” and the cottage industry which has sprung up around it.

When clinicians refer to narcissism, they are most often referring to narcissistic personality disorder (NPD), which is described in the DSM-V as meeting 5 of the 9 following criteria:

1. A grandiose sense of self-importance
2. A preoccupation with fantasies of unlimited success, power, brilliance, beauty, or ideal love
3. A belief that he or she is special and unique and can only be understood by, or should associate with, other special or high-status people or institutions
4. A need for excessive admiration
5. A sense of entitlement
6. Interpersonally exploitative behavior
7. A lack of empathy
8. Envy of others or a belief that others are envious of him or her
9. A demonstration of arrogant and haughty behaviors or attitudes

It should immediately become clear how tenuous this diagnosis is. Let’s take the first criterion: “a grandiose sense of self-importance.” Unlike the presence of a pathogen, there is no objective test for grandiosity; it depends entirely on clinician judgment. And clinicians often disagree. Some clinicians may even be grandiose themselves and have a distorted sense of what constitutes “real” grandiosity, missing otherwise obvious instances of NPD because they see themselves in their patients.

Let’s also consider “a sense of entitlement.” Entitlement here refers to someone who believes they deserve things they do not actually have a right to. But what one legitimately “has a right to” varies greatly between sociocultural groups. Consider a classic: are kids allowed to talk back to their parents when they feel their parents are being unfair? Your response most likely depends on the culture in which you were raised. A culture which places strong emphasis on filial piety may believe that any child who questions their parents is deeply entitled, and in these cultures such backtalk may be used as evidence to furnish an NPD diagnosis. In other cultures, children questioning their parents may be an acceptable or even expected part of life, and it may even be the accepted role of the child to keep the parent’s narcissism in check by calling their authority into question. It all depends on whether parental authority comes from the mere fact of being a parent, or being a just/good parent, and different cultures have different views.

Consider too that cultures are never distinct entities, and cultural ideals can interweave or conflict, even under the same roof. It’s more than plausible to imagine a situation in which a child questions a parent’s decision, and because the parent and the child have different cultural ideals (say, due to being raised in different places), each thinks the other is entitled. Throw in a clinician who shares one cultural idea but not the other, and you could end up with two very different diagnostic pictures.

These are just two examples of how fundamentally subjective the diagnostic criteria for NPD are. This isn’t even to mention that envy can be used to make the case that someone is a narcissist. I would be hard-pressed to find any clinician, indeed, any person I know, who did not consider envy to be a fundamental part of the human experience. Now, one could make the case that because the DSM requires five of the nine criteria, this lends the category of “narcissism” some validity and/or specificity. After all, not every envious person is a narcissist, but an envious person who also exhibits four of the other traits most certainly must be. Right?

Well, consider this: imagine a person who exhibits only four of the narcissistic traits, but in the most extreme manner. They believe themselves to be the center of the world and don’t care about others at all. Indeed, they are the most selfish and self-obsessed person you’ve ever met. But they only exhibit four of the nine narcissistic traits. Are they not a narcissist because they only meet four criteria instead of five? It wouldn’t make sense to put the 4/9 person in a different category from the 5/9 person if the 4/9 person is just as severe. But NPD makes this distinction, at least formally. (In practice, these two people might well be put in the same category, highlighting once again the role of clinician judgment in NPD diagnosis).

Thus we can see how vacant the category of “the narcissist” as a medical phenomenon actually is. Like many psychiatric categories, narcissism is diagnosed functionally, based on how much it interferes with a person’s life and the lives of those around them. Thus, when contemporary articles promise advice or life hacks to deal with “the narcissist” you find in your own life, be skeptical. There may be sound advice in those articles, but it almost certainly isn’t tailored to deal with “the narcissist” specifically, because “the narcissist” doesn’t really exist in the world. These articles often appeal to the pathologizing language of “narcissist” to lend themselves scientific/medical credibility (or specificity) they don’t actually have.

Academic psychology believes in the narcissist, and believes in ways to treat NPD. Reading peer-reviewed articles on responses to NPD behaviors is about as close as you could get to the lofty promise of “destroying” a narcissist. (Not to mention that any time you find yourself looking for a way to “destroy” someone, you have completely and utterly dehumanized them, and it may be worth reflecting on why that’s the case).

This is not to discredit, lessen, or dismiss the abuse that people have undergone at the hands of people who exhibit narcissistic traits, or to invalidate their rage at their abusers. Rather, I hope to highlight how hollow the category of “narcissist” is, even by academic standards, and to encourage critical discourse on the subject.

I recently learned academic astrology was (still) a thing. There’s even an entire journal devoted to the subject. Academic astrologers look for (occasionally statistical) correlations between celestial events and human personality traits and other psychological phenomena. The most common astrological practice is the production of a natal chart: based on the position of the planets and stars at one’s birth, astrologers believe that certain psychological traits will be present in them during adulthood.

Astrology is one of those weird subjects that, much like Jungian analysis, rides the line between a hermeneutic and a science: like a hermeneutic, astrology proposes an interpretation of events (and is largely uninterested in validating the basis of that interpretation). Like a science, astrology proposes a systematic interpretation, rather than one that is merely values-driven like other hermeneutics (historical-critical, feminist, post-structuralist).

In this essay, I want to do two things. First, I want to review arguments from the philosophy of science which thoroughly dismiss the scientific claims made by various schools of astrology and astrological thinkers. Second, I want to offer a rehabilitative reading of astrology that understands astrological practices as a cultural form of meaning-making, one whose value goes beyond simple prediction or explanation as an attempt to offer insight into the subjectivities of the people who practice it.

There are two main criticisms of astrology leveled by the philosophers of science. The first is that astrological predictions have been falsified on numerous occasions, which refutes astrology as a scientific theory. The second is that astrology lacks plausible mechanisms for its purported effects, and any proposed mechanisms contradict known laws of physics. Let us discuss each of these in turn.

Regarding falsification, a core tenet of any scientific theory is that it must be falsifiable, and, if proven false, must be discarded (Popper). Astrology makes predictions about the psychology of individuals based on the position of celestial bodies at their birth. If astrology is scientifically legitimate, these predictions should hold into adulthood. In 1985, an experiment was conducted which systematically compared astrologers’ assessments of psychological profiles based on natal charts with standardized psychological testing outcomes (in this case, the California Psychological Inventory). Even after controlling for astrologers’ confidence, astrologers failed to perform better than chance at predicting psychological profiles based on corresponding natal charts.

Similarly, a comprehensive 2003 review study finds that: “a large-scale test of persons born less than five minutes apart found no hint of the similarities predicted by astrology” and more significantly “meta-analysis of more than forty controlled studies suggests that astrologers are unable to perform significantly better than chance even on the more basic tasks such as predicting extraversion.” Thus, a review of the scientific literature finds robust grounds for rejecting the psychological claims made by astrologers.

The second claim, that astrology lacks plausible mechanisms for its purported phenomena, is more difficult to argue for. This is due to the fact that it is always easier to prove something exists than something does not exist. It is very challenging, to say for instance, that there is definitively no mechanism by which astrological alignments might work on human personality. It is more feasible to say that scientific experts regard every popular proposed mechanism (such as latent effects from the gravity of celestial bodies) as lacking theoretic plausibility. However, as with all science, the burden of proof lies with those making the claim, and astrology has yet to demonstrate in theoretic or empirical terms a mechanism by which someone’s natal chart has an effect on their psychological profile.

We thus have thorough reason to doubt in no uncertain terms the scientific status of astrology. The natural question then is why does an unscientific idea with no predictive power persist? Unlike some, I do not believe the persistence of astrology in America is merely due to a lack of scientific literacy among its (predominantly female) base. Indeed, the sexist notion that star charts are simply for women who don’t understand science pervades criticism of astrology. While there are certainly gaps in America’s public education system, especially in the sciences, a basic knowledge that human personality and world events are not determined by the alignment of celestial bodies is not one of them. The literature supports this, with a recent survey finding that just 30% of Americans “believe in” astrology. While this number seems large, consider that, relatively speaking, Americans understand the scientific status of astrology better than a range of other pseudoscientific phenomena. For example, 41% of Americans believe in psychics, and 53% believe in ghosts or spirit visitation.

Putting the complicated politics of “belief” and the intricacies of survey methodology aside, I believe a good portion of people who adhere to astrology in various ways are not engaged in unscientific thought, but rather produce meaning in their lives through dialogue with a collection of astrological ideas. Interpreting one’s life through natal astrology is markedly different from making high-stakes inferences or predictions based on it (as one does with legitimate belief).

Generally speaking, astrology is a form of archetypal storytelling. One reckons with the image of themselves as a Virgo much in the same way as with being told that “being hot-headed runs in the family.” While there may be a biological-psychological basis for actual or social inheritance of “hot-headedness,” it is fundamentally an (un)conscious relation to one’s prescribed archetype that creates the outcome. A child who is told that “being hot-headed runs in the family” may gain permission to be hot-headed, and consequently go through life conceiving of “hot-headedness” as a legitimate self-regulation paradigm. Conversely, a child may resist the notion that they are hot-headed and develop a counternarrative about themselves, such as “my parents are hot-headed, and I was supposed to be hot-headed like them, but I am calm.” This second person may go to greater lengths than average to produce or legitimize their “calmness” precisely because it runs counter to the archetypal narrative they were given about themselves as a child.

Regardless of whether one accepts or rejects their assigned archetypes, they serve as a way for people to tell stories about themselves and orient themselves in the world. This becomes relevant when we realize that astrology is nothing more than a collection of archetypes. Being “hot-headed” in alignment or opposition with one’s parents is no different than being neurotic in alignment or opposition to one’s status as a Virgo. “I’m a Virgo but I’m not neurotic” is just as prescriptive a statement as “I am a Virgo and I am neurotic:” both are statements of identity in response to an (arbitrary) preconceived idea.

Any time narratives are invoked in the creation of subjectivity, we have a culture, and understanding that culture is vital to truly reckoning with the subjectivities of the people who find themselves enmeshed within it. Offering an interpretation which takes up the figure of “the Virgo” to an astrologically-minded person may have far more use than offering a clinical-psychological interpretation of the same phenomena. Astrological archetypes are a language, and you have to be able to speak that language to understand the people who, for one reason or another, find it to be their native tongue.

Thus, while we can resoundingly dismiss the scientific claims of astrology, it is much harder to dismiss the value astrology may have as a tool of meaning-making and personal storytelling, especially for women and queer people who, for one reason or another, may have burned by traditional narratives of subjectivity, whether religious, scientific or otherwise.

1.) Specialization: Graduate school in any discipline requires hyper-specialization in a research topic which becomes the basis for one’s dissertation. When I was thinking about graduate school in comp lit, I fantasized about being immersed in the world’s literature, gaining an understanding of global thought and storytelling. When I realized that I would spend the majority of my time in a single body of literature, and probably selected works from one or two authors during a particular period, I became much less enthused. (So far, the only program I’ve found that seems close to what I thought comp lit grad school would be, namely a truly global literary training, is the Pacifica Graduate Institute’s M.A./Ph.D. in Mythological Studies, which has required or optional coursework in all the major world regions but is largely limited to myth and folktales rather than contemporary works).

2.) Lack of Proper Language Training: The undergraduate program I attended is probably the only comparative literature program in the United States that does not require advanced classes in a foreign language or its literature. To be competitive for doctoral programs in comp lit, an applicant usually needs mastery of at least two languages and a significant handle on a third by the end of their undergraduate career or master’s. I could have self-studied or made the case for myself, but by the time I was applying to programs, it was too late to gain the substantial language training I would need to be truly competitive. My peers who were successful in that process often came into undergrad with English and one other language, rigorously studied a high-demand language in college, and/or could passably read French and German by the time they graduated.

3.) ADHD: As I worked through undergrad, I struggled more and more with getting through longer and/or complex texts. I ended up getting diagnosed with ADHD during my junior year, which was a huge relief (I had begun to develop serious self-esteem issues related to my ability to focus through longer readings). While I could have powered through a doctoral program as I did the remainder of my undergrad (I thought for a time I might be able to pull it off with a good text-to-speech app) the prospect of spending years reading (or listening) to thousands of pages of text began to sound intolerable.

4.) Theorists became increasingly illegible: I got into comp lit because I love theories about how people and culture interact. It was authors like Michel Foucault, Walter Benjamin, and Susan Sontag that made me want to keep going. However, as I got deeper into the major, especially into upper-level theory electives, what the authors were saying became increasingly illegible, and I often doubted if they were saying anything at all. I could wave away the eccentricities of Derrida and Lacan (I naively thought “surely they must be better in French”) but the sheer unreadability of some of the texts I was assigned in my graduate courses was almost laughable. It didn’t help that my intended area of specialization was psychoanalytic theory. For context, in a graduate course I took on psychoanalytic subjectivity, we had to read the work of contemporary Lacanian psychoanalyst Jamieson Webster. Here’s an excerpt from the opening chapter of arguably her most famous book, Conversion Disorder:

“Nevertheless, knowledge and imagination do not occupy the same space. They are differently folded outside in or inside out. Imagination, in a sense, doesn’t look to occupy anything other than the field of attention, usually coming underfoot, at play, unoccupied. I cannot inhabit my imagination. Try as I may, I don’t have my imagination. Then who does? That is as good a place to begin as any, and it happens to be the place that psychoanalysis begins, my occupation of sorts” (32).

Jamieson Webster is one of the more intelligible people in this field; I genuinely understand (and like) some of her other work. But if this passage is saying anything at all, I’m not going the be the poor soul who has to puzzle it out.

5.) Demand for rigorous historiography and iterative scholarship: A major portion of becoming a scholar in any discipline is immersing yourself in the history of the field. That means reading everything about your area of specialization before you can decide what contribution you’re even able to make. A dissertation is fundamentally explaining how your work builds on (but is not derivative of) the work of your predecessors, and depending on the field/specialty, there can be a lot of predecessors which need to be properly discussed in your project. The thought of having to complete this task in any branch of psychoanalytic theory was overwhelming. I like reading texts and talking about them; the little historiographic work I did felt like asking for permission from 50+ years’ worth of scholars just to have the privilege of speaking. My philosophy (now) is: if I’ve read a text, I should be able to say what I want to about it, and I shouldn’t have to go back 50 years to figure out how what I’m saying fits into the existing scholarly discourse. Unfortunately, that is a bad philosophy to have as a graduate student in comparative literature.

I absolutely loved being a comp lit undergrad. A training in comparative literature (or any rigorous major in the humanities) teaches you how to think in a way few other things can. Nevertheless, I was happy to graduate from the program, as I had long begun to feel I was supposed to be somewhere else.

The film’s overt discussion of sexual control begins early on. At Lentulus Batiatus’ gladiatorial school, at which Spartacus is an enslaved student, those who “perform well”’ are rewarded with the “company” of a female companion in their cell. Spartacus is very quickly deemed worthy, and is “given” Varinia by the staff, at the special request of Batiatus himself. Spartacus refuses to initiate sex with Varinia, for which he is mocked by Batiatus and others, who are revealed to be watching from overhead. His captors claim that he will never be a man, to which Spartacus retorts by screaming “I am not an animal.”

This scene establishes efforts of Batiatus et al. to control the sexuality of their captives. Sex is not merely a reward for good behavior, rather we see early on that Batiatus et al. make deliberate (often arbitrary) decisions about which women to “mate” with which of the gladiators as a means of exerting control over the sexuality of both the men and the women at the gladiatorial school. It is made abundantly clear that there is not a random assignment of a certain number of women to a certain number of men; rather, Batiatus and his staff hand-pick each pairing. There is seemingly no functional purpose to this highly-individualized pairing process other than confirming for Batiatus and his staff their continued control over the sexual behavior of the enslaved. When Spartacus refuses to have sex with Varinia, he interrupts the narrative that Batiatus et al. have control over the sexual behavior of their captives, and Varinia is immediately removed from his cell. Once control over Spartacus’ sexuality is no longer a possibility, specifically once Batiatus et al. can no longer witness their control of Spartacus’ sexuality (they would seemingly have watched Spartacus have sex with Varinia had he initiated), the whole conjugal assignment process is immediately ceased. This confirms for us the nature of the program, which is not merely a psychological gambit to maintain power relations at the school in general, but also deeply invested in maintaining power over sexuality in particular, in this case when and with whom sexual arousal starts and stops.

Batiatus’ efforts to secure sexual control are underscored when it becomes clear to Batiatus et al. that Spartacus actually desires Varinia. Later on in the film, Varinia is escorted into Spartacus’ cell during the conjugate visit period, only to be immediately removed. This attempts to exert control over Spartacus’ desire, specifically by initiating arousal (introducing Varinia) and then interrupting it (removing her suddenly). This attempt at sexual control overlaps with Batiatus’ attempts to secure power more generally, as he signals his authority by showcasing his ability to grant and revoke objects of desire on a whim. Nevertheless, an analogous gesture could have been performed with the introduction and revocation of food or water, both of which could be argued have a more immediate bearing on Spartacus’ wellbeing. Batiatius chooses a sexual object for this maneuver because it performs the double function of demonstrating his authority over Spartacus’ body in general and his ability to fulfill his sexual desires. It finishes what was incomplete in Spartacus and Varinia’s initial introduction, namely demonstrating to Batiatus, if no one else, his control over the sexuality of those he enslaves.

Whereas Batiatus seeks control over the sexuality of his gladiators in a systematic sense (he cares not about an individual gladiator, but his ability to manipulate their desire as a class of people under his control), Crassus seeks sexual control over Antonitus in a personal sense. He seeks control not over the generalized class of “the slave” but over Antonitus in particular. To be sure, Crassus invokes Antonitus’ status as enslaved in attempting to coerce him into sex, but this is done to establish a power differential to facilitate particular sexual outcomes, not the other way around (as with Batiatus and the gladiators). In coded speech, Crassus asks Antonitus whether he prefers “snails or oysters” (penises or vaginas), and asks him further if he thinks it immoral to “eat snails.” Crassus then launches into a soliloquy about the all-conquering power of Rome, stating that the only thing “a boy” (Antonitus) can do in response to Rome’s demands is “serve…[and] love her” (submit to Crassus’ sexual requests). When Crassus finishes his speech, Antonitus is gone, having run away to (eventually) join Spartacus. Antonitus escaping from Crassus’ estate echoes the larger ideological gesture of Spartacus’ slave revolt: a refusal to submit to the power of the upper classes. However, the incident between Crassus and Antonitus localizes this refusal in the sexual. Antonitus leaves because he refuses to let Crassus use his status as slave to establish control over his sexuality.

It goes without saying that the writing of Spartacus is an overt critique of attempts to establish sexual control, either in a systematic or personal sense. Spartacus is the hero of the film, and it is expected the viewer will identify with his desire for liberation. The film’s “honeymoon” period is when Spartacus is free and he and Varinia are allowed to be with one another at their own discretion, in stark contrast to the rigid, modular encounters which occurred under Batiatus. While there is less direct commentary on Antonitus’ sexual freedom after his escape, Crassus is the closest thing Spartacus has to a villain, and Crassus’ actions (including coercing Antonitus) are meant to be understood as immoral.

Despite the writing of Spartacus having a clear investment in critiquing sexual control, the cinematography has a far more ambiguous relationship to forced intimacy, particularly between the viewer and the figures of Spartacus and Antonitus. In a training scene at the gladiatorial school, Spartacus is chosen to model the efficacy of various combat techniques. He stands at full attention while a staff member paints his body with red, blue, and yellow paint, with each denoting a particular region where a gladiator could strike. Red is “instant kill,” blue “get[s] a cripple” and yellow is a “slow kill.” Here, the viewer is not merely invited to ogle Spartacus’ (Kirk Douglas’) body, they are given a guided tour. The various regions of the body noteworthy for future gladiators also happen to be points of sensual interest: the lower neck, the inner arms, the abdomen. The scene covers highly vulnerable regions as well as “slow kill” regions, which pragmatically means the viewer gets a thorough encounter with almost the entirety of Spartacus’ upper body. Moreover, the paint used to mark these regions is excessively bright, almost neon, to the point where its color saturation is noticeably removed from the otherwise muted color palette of the film. The application of liquid to skin also heightens sensuality by arousing the sensation of touch: it could have been, for example, that someone simply pointed at various points on Spartacus’ body with a stick. But this is not the case: each region is caressed by a paint brush whose color is so overwhelming the viewer can hardly look at anything else on screen. The film is essentially directing the viewer’s (presumably sexual) gaze not just to Spartacus’ body, but particular regions of it, heightening perceived sensuality through the application of color and texture to the skin. Just as Batiatus seeks to watch his slaves engage in erotics he orchestrates, so the film seeks to direct the viewer through a kind of forced intimacy with the flesh of its titular protagonist.

A similar dynamic emerges when the film goes out of its way to showcase the body of Antonitus (Tony Curtis). The aforementioned “snails and oysters” scene begins with Crassus alone in a pool before he calls for Antonitus to assist him. Antonitus, wearing nothing but form-fitting underwear, slowly crosses the entirety of the screen to grab a stool before entering the pool. Now, the case can be made that the scene is attempting to show Crassus’ sexualization of Antonitus, and that the crossing of the screen is merely to give the viewer time to appreciate the extent to which Antonitus’ attire reveals his body. But Antonintus crossing the screen to grab a stool, an entirely superfluous choice given that the stool could have simply been placed on the other side of the screen, takes more than 10 seconds. For a full 10 seconds, there is nothing of note happening on screen except Antonitus walking about nearly nude. And once again we have the application of liquid to flesh to heighten sensuality: Crassus could have asked for Antonitus to massage him anywhere, but his asking for it in a pool means that Antonitus’ body glistens while he works. This heightened visual sensuality is presumably not for Crassus either, as he faces away from Antonitus for almost the entirety of the scene; rather, the viewer alone is asked to stare at Antonitus’ torso and be intimate with it as he becomes the only dynamic figure on screen.

This overt display of Antonitus’ body is underscored in the next scene, when the two get out of the pool: Crassus wraps his body, Antonitus does not. For another full 10 seconds, the viewer is permitted to gaze at Antonitus’ still wet torso and thighs in the background while Crassus walks slowly towards the camera. Once again, the film undercuts its writing by facilitating forced intimacies between the viewer and the flesh of the film’s enslaved characters, practicing the sexualized control of enslaved bodies which Spartacus explicitly revolts against.

Spartacus is fundamentally about a revolution which outlives its material ends. A key part of that revolution is sexual: enslaved people retaking control over their bodies from their Roman captors. However, in the process of portraying that revolution, the film also demands its viewer be sensually intimate with those bodies that are seeking freedom. It asks the filmgoer to not merely consider enslaved people as sexual subjects, but orchestrates the viewer’s sexualization of them. Thus, the film’s overt goal of critiquing sexual control is undercut by its insistence on exerting that sexual control for some thematic, economic (or perhaps unconscious) end. A more radical reading might be that Spartacus is actually disinterested in critiquing sexual control, and instead seeks to inspire in the viewer through its cinematography a kind of Roman sexuality: a sexuality which looks across power differentials for arousal, that seeks to actually be Batiatus or Crassus rather than overthrow them. This would deeply complicate the film by imagining the cinematography not merely as undercutting the writing, but acting (perhaps unconsciously) as its antagonist. Ultimately, no film is capable of speaking in a unified voice, and Spartacus merits more thorough analysis than can be performed here. Still, I hope to have inspired in the reader a consideration of the ways in which reading the sexual aspects of Spartacus is more complicated than it may initially appear.